Re: [tied] Re: Derivations 15 sun

On Mon, 18 Oct 2004 11:56:18 +0200, Piotr Gasiorowski
<gpiotr@...> wrote:

>
>On 04-10-18 10:36, Miguel Carrasquer wrote:
>
>> Alb. hyl (*suh2li- or
>> *suh2lo-), all with metathesis *h2u > *uh2, ...
>
>If it belongs here at all. There's more chance that <diell> 'sun' does
>(< *s(h2)wél...). <yll> ~ <hyll> 'star' (not 'sun', and with the /h-/ ~
>zero fluctuation even between neighbouring dialects) is, in my opinion,
>easier to connect with *h1us- (cf. OE ysel, ysle 'spark, ember', ON
>ysli, MHG usele < PGmc. *us-il-o:/-a[+ *-n-]). The vowel lengthening in
>*h1us-lo- > *u:ll- > (h)yll may be a regular compensatory change, see a
>similar Albanian lengthening before *-rs-/*-sr- > *-:r-.

You're right, this has been discussed here before. I was
merely enumerating the cognate list from Pokorny.

Now <diell> is of course a much more interesting form,
confirming as it does the root-shape *sh2wé(l/n)-, as also
found in Avestan <xv&:N>.

=======================
Miguel Carrasquer Vidal
mcv@...